Factors associated with COVID-19 vaccination among pregnant … – Nature.com
October 25, 2023
Study setting, study design, and population
This retrospective cohort study used linked data from the live birth and vaccination information systems in Rio de Janeiro city. Rio de Janeiro is the second largest city in Brazil, with an estimated population of 6,700,000. It has a monthly per capita income of 4 minimum wage, considered a high-income city compared to other regions of the country37 In 2021, the number of live births was 67,97338. In July 2022, the coverage of COVID-19 vaccination was 89% for the complete primary regimen and 57% for a booster dose in the general population above 18 years39.
The COVID-19 immunisation started in Rio de Janeiro city on January 20, 202140, initially for healthcare professionals, adults older than 60 years, and high-risk groups. Pregnant women with comorbidities started vaccination in March 2021; by May 2021, it was interrupted due to a severe adverse event related to the COVID-19 vaccine in this population. From July 07, 2021, vaccination was resumed and made available for all pregnant, breastfeeding or women planning to become pregnant41. CoronaVac (Sinovac/Butantan) and BNT162b2 (Pfizer/Biotech) were the recommended platforms.
For this investigation, we included all women who delivered live births in Rio de Janeiro city between 1st August 2021 and 31 July 2022. Our exclusion criteria were multiple births, ages less than 18 years or higher than 49 years, and records with missing or implausible gestational ages (> 44 weeks) (Fig. 2).
The Declaration of Live Birth, a legal document filled out by the health care professional who attends the birth, is entered into the Live Birth Information System (SINASC). It contains details on the mother (such as age, education, skin colour, and marital status), about the pregnancy (such as antenatal appointments, the gestational period, prior gestations, previous live births, and previous losses), and details about the newborn (e.g., birth weight, sex, APGAR score). In addition, all vaccinations provided in Brazil are documented in the National Immunization Program Information System (SI-PNI), along with the administration date of the first, second, and booster doses, with its platform type. The SINASC data initially available had records from women who gave a live birth in Rio de Janeiro city from January 1, 2020, to August 28, 2022. The SI-PNI data included vaccination records from January 19, 2021, to August 31, 2022 (Supplementary Figure 1, Table 1 and Table 2). The linkage between records from SI-PNI and SINASC allowed access to any vaccination that happened before, during and after the pregnancy period.
The matching process used the maternal name, date of birth, zip code, and neighbourhood. We used the Jaro-Winkler string comparator to compare the similarity between string variables recorded in SINASC and SI-PNI. This algorithm calculates the similarity between two strings based on the number of shared characters and transpositions42. The resulting similarity score ranges between 0 (no similarity) and 1 (perfect similarity). We categorised the similarity scores into three categories: (0, 0.85), (0.85, 0.95), and (0.95, 1)]. We employed a three-step approach that checked for a string similarity score greater than 0.95, followed by exact matches for dates of birth and zip code. Any potential matches then underwent a manual review. After data linkage, the individual identifiers were removed, and the de-identified dataset was made available for analysis.
We estimated the date of the last period (DLP) using the date of delivery minus the days of pregnancy according to the gestational age at birth. We determined the date of conception by adding 14 days to the DLP. We defined the gestational period as the time between the date of conception and the date of birth. The vaccination status was determined using the dates of the vaccination registries compared to the gestational period.
Women who received at least one dose between the conception date and the date of delivery were considered vaccinated during pregnancy. Those who received all registered doses before the pregnancy period (before the conception date) were grouped as vaccinated before pregnancy. Those who received vaccines exclusively after the delivery date were assigned as vaccinated after pregnancy. Finally, women with no register of a vaccine dose were regarded as never vaccinated. We estimated vaccine uptake during pregnancy as the proportion of women who received any vaccination during pregnancy as a percentage of all births (Supplementary Table 3).
We divided variables into sociodemographic: age (1824, 2534, 35 years), education (07, 811, 12 years), self-identified skin colour (black, parda/brown, white, Asian, or Indigenous), and marital status (with or without a partner). And obstetric: the trimester of the first antenatal care appointment (first, second, or third), the total number of appointments (03, 46, or >6), the number of previous gestations (none, 12, or 3), the number of previous live births (none, 12, or 3), and the number of previous child loss (none or at least one). The Indigenous and Asian races were presented in the descriptive analysis and excluded from the logistic regression due to the small sample size.
We considered the length of pregnancy and the burden of COVID-19 infection during the study period to be a confounder a priori. Therefore, we included the month-year of birth and the gestational week at delivery as additional variables in the analyses.
We assessed each groups characteristics by describing categorical variables as frequencies and percentages, excluding missing data. Continuous variables, such as age and gestational age, were presented as the median and interquartile range (IQR). In the descriptive analysis, we stratified the groups by being vaccinated only before pregnancy, vaccinated with at least one dose during pregnancy, and never vaccinated.
To identify the factors associated with vaccine uptake in pregnant women, we compared only the population of women vaccinated during pregnancy with those who were never vaccinated during the study period. For each potential factor associated with uptake, we ran a bivariate logistic regression individually, describing the crude odds ratio (OR) and its associated 95% confidence interval, controlled by gestational age and month-year of birth (Supplementary Table 7).
In addition, we performed a multiple logistic regression using a hierarchical framework with two levels. In the first level, we had the socioeconomic variables: age, education, self-identified skin colour, and marital status. In the second level, we included all the variables above and the obstetrics variables: the total number of antenatal appointments, the number of previous live births, and the number of previous child losses.
In the first model of multiple logistic regression, we included only the socioeconomic variables. The overall effect of socioeconomic factors (the distal factors) was assessed in this model 1. In the second model, we included the obstetric variables in addition to the sociodemographic block. Therefore, the unconfounded effect of the obstetric variables was obtained in this model. Both models were also controlled by gestational age and month-year of birth. Missing data on each covariate were excluded from the analysis (Supplementary Table 8). The final adjusted odds ratio and 95% confidence intervals were described for each model separately (Supplementary Table 7). Data management and statistical analysis were conducted using IBM SPSS, Statistical Package for the Social Sciences, Version 28.0 (Armonk, NY: IBM Corp). The de-identified dataset and the programming codes can be made available under the request.
The present study has been approved by the Ethics Committee of Research Center Gonalo Moniz /Oswaldo Cruz Foundation, Salvador, Bahia, Brazil (IORG 0002090/OMB No. 0990-0279 valid until 01/27/2025), under the Certificate of Submission for Ethics Review No 63287822.0.0000.0040. The protocol and procedures presented in the project are in full accordance with the Brazilian legislation (Resolution CNS 466/2012) and the declaration of Helsinki regarding ethical standards in conducting research involving human beings. Due to the retrospective nature of the study, the need for informed consent was waived by the Ethics Committee of Research Center Gonalo Moniz /FIOCRUZ/BA.
See the original post:
Factors associated with COVID-19 vaccination among pregnant ... - Nature.com
